An empirical approach to characterizing trunk muscle coactivation using simulation input modeling techniques.

Accurately describing trunk muscle coactivation is fundamental to quantifying the spine reaction forces that occur during lifting tasks and has been the focus of a great deal of research in the spine biomechanics literature. One limitation of previous approaches has been a lack of consideration given to the variability in these coactivation strategies. The research presented in this paper is an empirical approach to quantifying and modeling trunk muscle coactivation using simulation input modeling techniques. Electromyographic (EMG) data were collected from 28 human subjects as they performed controlled trunk extension exertions. These exertions included isokinetic (10 and 45 degrees /s) and constant acceleration (50 degrees /s/s) trunk extensions in symmetric and asymmetric (30 degrees ) postures at two levels of trunk extension moment (30 and 80Nm). The EMG data were collected from the right and left pairs of the erector spinae, latissimus dorsi, rectus abdominis, external obliques and internal obliques. Each subject performed nine repetitions of each combination of independent variables. The data collected during these trials were used to develop marginal distributions of trunk muscle activity as well as a 10x10 correlation matrix that described how the muscles cooperated to produce these extension torques. These elements were then combined to generate multivariate distributions describing the coactivation of the trunk musculature. An analysis of these distributions revealed that increases in extension moment, extension velocity and sagittal flexion angle created increases in both the mean and the variance of the distributions of the muscular response, while increases in the rate of trunk extension acceleration decreased both the mean and variance of the distributions of activity across all muscles considered. Increases in trunk asymmetry created a decrease in mean of the ipsi-lateral erector spinae and an increase in the mean of all other muscles considered, but there was little change in the variance of these distributions as a function of asymmetry.